add support for dicom images (#1050)

Signed-off-by: Can Zhao <[email protected]>

Fixes #.

### Description
The original source of mhd/raw images was removed. Now add support for
dicom images.
And reformatted code.

### Checks
<!--- Put an `x` in all the boxes that apply, and remove the not
applicable items -->
- [ ] Notebook runs automatically `./runner [-p <regex_pattern>]`

Signed-off-by: Can Zhao <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Author: Can-Zhao

detection/README.md

@@ -19,6 +19,8 @@ The dataset we are experimenting in this example is LUNA16 (https://luna16.grand
 
 LUNA16 is a public dataset of CT lung nodule detection. Using raw CT scans, the goal is to identify locations of possible nodules, and to assign a probability for being a nodule to each location.
 
+Users can either download mhd/raw data from [LUNA16](https://luna16.grand-challenge.org/Home/), or DICOM data from [LIDC-IDRI](https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=1966254).
+
 Disclaimer: We are not the host of the data. Please make sure to read the requirements and usage policies of the data and give credit to the authors of the dataset! We acknowledge the National Cancer Institute and the Foundation for the National Institutes of Health, and their critical role in the creation of the free publicly available LIDC/IDRI Database used in this study.
 
 We follow the official 10-fold data splitting from LUNA16 challenge and generate data split json files using the script from [nnDetection](https://github.com/MIC-DKFZ/nnDetection/blob/main/projects/Task016_Luna/scripts/prepare.py).
@@ -36,15 +38,21 @@ In these files, the values of "box" are the ground truth boxes in world coordina
 
 The raw CT images in LUNA16 have various of voxel sizes. The first step is to resample them to the same voxel size, which is defined in the value of "spacing" in [./config/config_train_luna16_16g.json](./config/config_train_luna16_16g.json).
 
-Then, please open [luna16_prepare_env_files.py](luna16_prepare_env_files.py), change the value of "raw_data_base_dir" to the directory where you store the downloaded images, the value of "downloaded_datasplit_dir" to where you downloaded the data split json files, and the value of "resampled_data_base_dir" to the target directory where you will save the resampled images.
+Then, please open [luna16_prepare_env_files.py](luna16_prepare_env_files.py), change the value of "raw_data_base_dir" to the directory where you store the downloaded images, the value of "downloaded_datasplit_dir" to where you downloaded the data split json files, and the value of "resampled_data_base_dir" to the target directory where you will save the resampled images. If you are using DICOM data, please also provide path to "dicom_meta_data_csv" which can be found in the downloaded folder from LIDC-IDRI.
 
-Finally, resample the images by running
+If you downloaded mhd/raw data, please resample the images by running
 ```bash
 python3 luna16_prepare_env_files.py
 python3 luna16_prepare_images.py -c ./config/config_train_luna16_16g.json
 ```
 
-The original images are with mhd/raw format, the resampled images will be with Nifti format.
+If you downloaded DICOM data, please resample the images by running
+```bash
+python3 luna16_prepare_env_files.py
+python3 luna16_prepare_images_dicom.py -c ./config/config_train_luna16_16g.json
+```
+
+The resampled images will be with Nifti format.
 
 #### [3.2 3D Detection Training](./luna16_training.py)
 

detection/generate_transforms.py

@@ -25,9 +25,21 @@
     RandFlipBoxd,
     RandRotateBox90d,
     RandZoomBoxd,
-    ConvertBoxModed
+    ConvertBoxModed,
 )
-def generate_detection_train_transform(image_key, box_key, label_key, gt_box_mode, intensity_transform, patch_size, batch_size, affine_lps_to_ras=False, amp=True):
+
+
+def generate_detection_train_transform(
+    image_key,
+    box_key,
+    label_key,
+    gt_box_mode,
+    intensity_transform,
+    patch_size,
+    batch_size,
+    affine_lps_to_ras=False,
+    amp=True,
+):
     """
     Generate training transform for detection.
 
@@ -166,7 +178,16 @@ def generate_detection_train_transform(image_key, box_key, label_key, gt_box_mod
     )
     return train_transforms
 
-def generate_detection_val_transform(image_key, box_key, label_key, gt_box_mode, intensity_transform, affine_lps_to_ras=False, amp=True):
+
+def generate_detection_val_transform(
+    image_key,
+    box_key,
+    label_key,
+    gt_box_mode,
+    intensity_transform,
+    affine_lps_to_ras=False,
+    amp=True,
+):
     """
     Generate validation transform for detection.
 
@@ -211,7 +232,17 @@ def generate_detection_val_transform(image_key, box_key, label_key, gt_box_mode,
     )
     return val_transforms
 
-def generate_detection_inference_transform(image_key, pred_box_key, pred_label_key, pred_score_key, gt_box_mode, intensity_transform, affine_lps_to_ras=False, amp=True):
+
+def generate_detection_inference_transform(
+    image_key,
+    pred_box_key,
+    pred_label_key,
+    pred_score_key,
+    gt_box_mode,
+    intensity_transform,
+    affine_lps_to_ras=False,
+    amp=True,
+):
     """
     Generate validation transform for detection.
 
@@ -260,7 +291,9 @@ def generate_detection_inference_transform(image_key, pred_box_key, pred_label_k
                 image_meta_key_postfix="meta_dict",
                 affine_lps_to_ras=affine_lps_to_ras,
             ),
-            ConvertBoxModed(box_keys=[pred_box_key], src_mode = "xyzxyz", dst_mode=gt_box_mode),
+            ConvertBoxModed(
+                box_keys=[pred_box_key], src_mode="xyzxyz", dst_mode=gt_box_mode
+            ),
             DeleteItemsd(keys=[image_key]),
         ]
     )

detection/luna16_post_combine_cross_fold_results.py

@@ -3,12 +3,14 @@
 import argparse
 import os
 
+
 def main():
     parser = argparse.ArgumentParser(description="PyTorch Object Detection Training")
     parser.add_argument(
         "-i",
         "--input",
-        nargs='+', default=[],
+        nargs="+",
+        default=[],
         help="input json",
     )
     parser.add_argument(
@@ -22,16 +24,20 @@ def main():
     in_json_list = args.input
     out_csv = args.output
 
-    with open(out_csv, 'w', newline='') as csvfile:
-        spamwriter = csv.writer(csvfile, delimiter=',',
-                                quotechar='|', quoting=csv.QUOTE_MINIMAL)
-        spamwriter.writerow(['seriesuid','coordX','coordY','coordZ','probability'])
+    with open(out_csv, "w", newline="") as csvfile:
+        spamwriter = csv.writer(
+            csvfile, delimiter=",", quotechar="|", quoting=csv.QUOTE_MINIMAL
+        )
+        spamwriter.writerow(["seriesuid", "coordX", "coordY", "coordZ", "probability"])
         for in_json in in_json_list:
             result = json.load(open(in_json, "r"))
             for subj in result["validation"]:
                 seriesuid = os.path.split(subj["image"])[-1][:-7]
                 for b in range(len(subj["box"])):
-                    spamwriter.writerow([seriesuid]+subj["box"][b][0:3]+[subj["score"][b]])
+                    spamwriter.writerow(
+                        [seriesuid] + subj["box"][b][0:3] + [subj["score"][b]]
+                    )
+
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     main()

detection/luna16_prepare_env_files.py

@@ -14,15 +14,31 @@
 import logging
 import sys
 
+
 def main():
     #  ------------- Modification starts -------------
-    raw_data_base_dir = "/orig_datasets/"  # the directory of the raw images
-    resampled_data_base_dir = "/datasets/"  # the directory of the resampled images
-    downloaded_datasplit_dir = "LUNA16_datasplit"  # the directory of downloaded data split files
+    raw_data_base_dir = "/home/canz/Projects/datasets/LIDC/manifest-1600709154662/LIDC-IDRI"  # the directory of the raw images
+    resampled_data_base_dir = "/home/canz/Projects/datasets/LIDC/manifest-1600709154662/LIDC-IDRI_resample"  # the directory of the resampled images
+    downloaded_datasplit_dir = (
+        "./LUNA16_datasplit"  # the directory of downloaded data split files
+    )
+
+    out_trained_models_dir = (
+        "trained_models"  # the directory to save trained model weights
+    )
+    out_tensorboard_events_dir = (
+        "tfevent_train"  # the directory to save tensorboard training curves
+    )
+    out_inference_result_dir = (
+        "result"  # the directory to save predicted boxes for inference
+    )
 
-    out_trained_models_dir = "trained_models"  # the directory to save trained model weights
-    out_tensorboard_events_dir = "tfevent_train"  # the directory to save tensorboard training curves
-    out_inference_result_dir = "result"  # the directory to save predicted boxes for inference
+    # if deal with mhd/raw data, set it to be None
+    # dicom_meta_data_csv = None
+    # if deal with DICOM data, also need metadata.csv
+    dicom_meta_data_csv = (
+        "/home/canz/Projects/datasets/LIDC/manifest-1600709154662/metadata.csv"
+    )
     #  ------------- Modification ends ---------------
 
     try:
@@ -45,20 +61,36 @@ def main():
     env_dict = {}
     env_dict["orig_data_base_dir"] = raw_data_base_dir
     env_dict["data_base_dir"] = resampled_data_base_dir
-    env_dict["data_list_file_path"] = os.path.join(downloaded_datasplit_dir,"original/dataset_fold0.json")
+    if dicom_meta_data_csv != None:
+        env_dict["data_list_file_path"] = os.path.join(
+            downloaded_datasplit_dir, "dicom_original/dataset_fold0.json"
+        )
+    else:
+        env_dict["data_list_file_path"] = os.path.join(
+            downloaded_datasplit_dir, "mhd_original/dataset_fold0.json"
+        )
+    if dicom_meta_data_csv != None:
+        env_dict["dicom_meta_data_csv"] = dicom_meta_data_csv
     with open(out_file, "w") as outfile:
         json.dump(env_dict, outfile, indent=4)
 
-
     # generate env json file for training and inference
     for fold in range(10):
-        out_file = "config/environment_luna16_fold"+str(fold)+".json"
+        out_file = "config/environment_luna16_fold" + str(fold) + ".json"
         env_dict = {}
-        env_dict["model_path"] = os.path.join(out_trained_models_dir,"model_luna16_fold"+str(fold)+".pt")
+        env_dict["model_path"] = os.path.join(
+            out_trained_models_dir, "model_luna16_fold" + str(fold) + ".pt"
+        )
         env_dict["data_base_dir"] = resampled_data_base_dir
-        env_dict["data_list_file_path"] = os.path.join(downloaded_datasplit_dir,"dataset_fold"+str(fold)+".json")
-        env_dict["tfevent_path"] = os.path.join(out_tensorboard_events_dir,"luna16_fold"+str(fold))
-        env_dict["result_list_file_path"] = os.path.join(out_inference_result_dir,"result_luna16_fold"+str(fold)+".json")
+        env_dict["data_list_file_path"] = os.path.join(
+            downloaded_datasplit_dir, "dataset_fold" + str(fold) + ".json"
+        )
+        env_dict["tfevent_path"] = os.path.join(
+            out_tensorboard_events_dir, "luna16_fold" + str(fold)
+        )
+        env_dict["result_list_file_path"] = os.path.join(
+            out_inference_result_dir, "result_luna16_fold" + str(fold) + ".json"
+        )
         with open(out_file, "w") as outfile:
             json.dump(env_dict, outfile, indent=4)
 

detection/luna16_prepare_images_dicom.py

@@ -0,0 +1,132 @@
+# Copyright (c) MONAI Consortium
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#     http://www.apache.org/licenses/LICENSE-2.0
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import argparse
+import json
+import logging
+import sys
+import os
+import csv
+from pathlib import Path
+
+import monai
+import torch
+from monai.data import DataLoader, Dataset, load_decathlon_datalist, NibabelWriter
+from monai.data.utils import no_collation
+from monai.transforms import (
+    Compose,
+    EnsureChannelFirstd,
+    EnsureTyped,
+    LoadImaged,
+    Orientationd,
+    Spacingd,
+)
+
+
+def main():
+    parser = argparse.ArgumentParser(description="LUNA16 Detection Image Resampling")
+    parser.add_argument(
+        "-e",
+        "--environment-file",
+        default="./config/environment_luna16_prepare.json",
+        help="environment json file that stores environment path",
+    )
+    parser.add_argument(
+        "-c",
+        "--config-file",
+        default="./config/config_train_luna16_16g.json",
+        help="config json file that stores hyper-parameters",
+    )
+    args = parser.parse_args()
+
+    monai.config.print_config()
+
+    env_dict = json.load(open(args.environment_file, "r"))
+    config_dict = json.load(open(args.config_file, "r"))
+
+    for k, v in env_dict.items():
+        setattr(args, k, v)
+    for k, v in config_dict.items():
+        setattr(args, k, v)
+
+    # 1. define transform
+    # resample images to args.spacing defined in args.config_file.
+    process_transforms = Compose(
+        [
+            LoadImaged(
+                keys=["image"],
+                meta_key_postfix="meta_dict",
+                reader="itkreader",
+                affine_lps_to_ras=True,
+            ),
+            EnsureChannelFirstd(keys=["image"]),
+            EnsureTyped(keys=["image"], dtype=torch.float16),
+            Orientationd(keys=["image"], axcodes="RAS"),
+            Spacingd(keys=["image"], pixdim=args.spacing, padding_mode="border"),
+        ]
+    )
+
+    # 2. prepare data
+    meta_dict = {}
+    with open(env_dict["dicom_meta_data_csv"], newline="") as csvfile:
+        print("open " + env_dict["dicom_meta_data_csv"])
+        reader = csv.DictReader(csvfile)
+        for row in reader:
+            meta_dict[row["File Location"][12:]] = str(row["Series UID"])
+
+    for data_list_key in ["training", "validation"]:
+        # create a data loader
+        process_data = load_decathlon_datalist(
+            args.data_list_file_path,
+            is_segmentation=True,
+            data_list_key=data_list_key,
+            base_dir=args.orig_data_base_dir,
+        )
+        process_ds = Dataset(
+            data=process_data,
+            transform=process_transforms,
+        )
+        process_loader = DataLoader(
+            process_ds,
+            batch_size=1,
+            shuffle=False,
+            num_workers=1,
+            pin_memory=False,
+            collate_fn=no_collation,
+        )
+
+        print("-" * 10)
+        for batch_data in process_loader:
+            for batch_data_i in batch_data:
+                subj_id = meta_dict[
+                    "/".join(
+                        batch_data_i["image_meta_dict"]["filename_or_obj"].split("/")[
+                            -3:
+                        ]
+                    )
+                ]
+                new_path = os.path.join(args.data_base_dir, subj_id)
+                Path(new_path).mkdir(parents=True, exist_ok=True)
+                new_filename = os.path.join(new_path, subj_id + ".nii.gz")
+                writer = NibabelWriter()
+                writer.set_data_array(data_array=batch_data_i["image"])
+                writer.set_metadata(meta_dict=batch_data_i["image"].meta)
+                writer.write(new_filename, verbose=True)
+
+
+if __name__ == "__main__":
+    logging.basicConfig(
+        stream=sys.stdout,
+        level=logging.INFO,
+        format="[%(asctime)s.%(msecs)03d][%(levelname)5s](%(name)s) - %(message)s",
+        datefmt="%Y-%m-%d %H:%M:%S",
+    )
+    main()